中文摘要：

针对基本混合蛙跳算法的收敛速度慢、容易陷入局部最优的缺点，提出了一种基于细胞通信策略的改进算法，该算法通过修改更新策略，从而增加了种群的多样性，产生更多靠近优质解的个体。用典型测试函数对基本蛙跳算法和改进的蛙跳算法及其他算法进行对比实验，仿真结果表明改进的蛙跳算法能较大幅度提高收敛精度。将改进的蛙跳算法应用于碳纤维生产过程水浴牵伸控制系统的优化，仿真结果表明其具有较好的优化控制效果。

英文摘要：

Basic shuffled frog leaping algorithm （SFLA） easily traps into local minimum and has slower convergence. To overcome these shortcomings, this paper proposes an improved SFI.A, which combines the cell communication mechanism. By modifying the update formula, the improved algorithm can maintain the population diversity and enhance the convergence velocity and precision. The comparison with several kinds of optimization algorithms are made for four benchmark test functions. Experimental results show that the improved SFLA is effective and robust. Finally, the improved algorithm is applied to the optimization of the water bath stretching slot control system in carbon fiber production, which further demonstrates the effectiveness of the improved SFLA.